Origin, Composition, and Structure of the Supernumerary B Chromosome of Drosophila

Genetics: Early Online, published on September 24, 2018 as 10.1534/genetics.118.301478

Page 1

1 Origin, composition, and structure of the supernumerary B chromosome of Drosophila

2 melanogaster

3

* † * *,‡ 4 Stacey L. Hanlon , Danny E. Miller , Salam Eche , and R. Scott Hawley

5

* 6 Stowers Institute for Medical Research, Kansas City, Missouri, 64110

† 7 Department of Pediatrics, Seattle Children's Hospital and University of Washington, Seattle,

8 Washington, 98105

‡ 9 Department of Molecular and Integrative Physiology, University of Kansas Medical Center, Kansas

10 City, KS, 66160

Copyright 2018. Page 2

11 Running title: D. melanogaster B chromosomes

12

13 Key words: B chromosome, isochromosome, centromere misdivision, satellite DNA, AAGAT

14

15 Corresponding author: Stacey L. Hanlon

16 Stowers Institute for Medical Research

th 17 1000 East 50 Street, Kansas City, MO 64110

18 Phone: 816-926-4146

19 E-mail: [email protected] Page 3

20 ABSTRACT

21 The number of chromosomes carried by an individual species is one of its defining characteristics. Some

22 species, however, can also carry supernumerary chromosomes referred to as B chromosomes. B

23 chromosomes were recently identified in a laboratory stock of Drosophila melanogaster—an established

24 model organism with a wealth of genetic and genomic resources—enabling us to subject them to

25 extensive molecular analysis. We isolated the B chromosomes by pulsed-field gel electrophoresis and

26 determined their composition through next-generation sequencing. Although these B chromosomes

27 carry no known euchromatic sequence, they are rich in transposable elements and long arrays of short

28 nucleotide repeats, the most abundant being the uncharacterized AAGAT satellite repeat. Fluorescent in-

29 situ hybridization on metaphase chromosome spreads revealed this repeat is located on Chromosome 4,

30 strongly suggesting the origin of the B chromosomes is Chromosome 4. Cytological and quantitative

31 comparisons of signal intensity between Chromosome 4 and the B chromosomes supports the hypothesis

32 that the structure of the B chromosome is an isochromosome. We also report the identification of a new

33 B chromosome variant in a related laboratory stock. This B chromosome has a similar repeat signature

34 as the original but is smaller and much less prevalent. We examined additional stocks with similar

35 genotypes and did not find B chromosomes, but did find these stocks lacked the AAGAT satellite repeat.

36 Our molecular characterization of D. melanogaster B chromosomes is the first step towards

37 understanding how supernumerary chromosomes arise from essential chromosomes and what may be

38 necessary for their stable inheritance. Page 4

39 INTRODUCTION

40 The growth, development, and reproduction of an organism relies on genetic material that is organized

41 into chromosomes. The chromosome complement carried by all members of a species is comprised of

42 essential chromosomes referred to as the “A chromosomes.” A subset of individuals within a species

43 may also possess extra chromosomes that are nonessential and not members of the standard A

44 chromosome set. These supernumerary chromosomes are commonly referred to as “B chromosomes.”

45 First described in 1907, B chromosomes have now been identified in hundreds of species across many

46 different taxa, and it is estimated that B chromosomes may be present in 15% of all eukaryotic species

47 (Wilson 1907; Jones and Rees 1982; Camacho 2005; D’Ambrosio et al. 2017).

48 The B chromosomes found in maize (Zea mays) are perhaps the best understood to date, and it is

49 from their presence in corn that the name “Type-B chromosome” originated (Randolph 1928). The

50 gravitation to study B chromosomes in corn was bolstered in part due to the advanced genetic and

51 cytological tools available at the time, and B chromosomes have since served as useful tools to study

52 and manipulate the genome (Beckett 1991; Birchler 1991; Yu et al. 2007; Masonbrink and Birchler

53 2012; Han et al. 2018). The recent rise of genomic analysis has allowed B chromosomes from several

54 species to be examined on a molecular level, revealing more about their potential origin and genetic

55 composition (Banaei-Moghaddam et al. 2015; Valente et al. 2017; Ruban et al. 2017). For instance, the

56 sequencing of rye (Secale cereal) B chromosomes demonstrated they arose just over 1 million years ago

57 and are derived from fragments of multiple A chromosomes (Martis et al. 2012). The B chromosomes in

58 the cichlid fish Astatotilapia latifasciata were sequenced and found to carry euchromatic sequence—

59 some of which is transcriptionally active—that is also present on several of the A chromosomes

60 (Valente et al. 2014). More recently, sequencing the B chromosome of the grasshopper Eumigus

61 monticola indicated that while it likely arose from one of the A chromosomes, it has since undergone

62 amplification (Ruiz-Ruano et al. 2017), and deep-sequencing in another grasshopper, Eyprepocnemis Page 5

63 plorans, revealed the presence of 10 protein-coding genes carried on the B chromosome (Navarro-

64 Domínguez et al. 2017). These studies demonstrate that the diversity and complexity of supernumerary

65 B chromosomes has been underestimated, and a better understanding of their origin and composition is

66 necessary.

67 Further work on the molecular characterization of B chromosomes has been impeded by several

68 obstacles. First, the majority of identified B chromosomes have only been studied cytologically in

69 samples collected from wild populations. Many of these species are not conducive to husbandry in the

70 lab, precluding development of the molecular, genetic, and genomic tools necessary to allow in-depth

71 study of B chromosomes. Second, the inability to control breeding and environment has the potential to

72 introduce uncontrolled genetic variation and unknown pressure on B chromosome evolution, leading to

73 fluctuations in B chromosome variants, frequency, and transmission rates (Zurita et al. 1998; Araújo et

74 al. 2002; Bakkali and Camacho 2004; Manrique-Poyato et al. 2013; Lanzas et al. 2018). Third, it is

75 challenging to precisely determine the age of a B chromosome in these wild systems, making it difficult

76 to discern how they formed and what events led to their current composition. Thus, as pointed out by

77 Jones (1995), B chromosome observations are only able to reflect the system in the present and therefore

78 may not reflect how the system was in the past.

79 Recently, B chromosomes were identified in a laboratory stock of Drosophila melanogaster

80 (Bauerly et al. 2014). Although B chromosomes have previously been observed in a handful of wild

81 populations within the Drosophila genus [albomicans (Clyde 1980); malerkotliana (Tonomura and

82 Tobari 1983); kikkawai (Sundaran and Gupta 1994); subsilvestris (Gutknecht et al. 1995); lini and

83 pseudoananassae (Deng et al. 2007)], Bauerly et al. (2014) was the first report of B chromosomes in

84 melanogaster. The laboratory stock in which these were identified carries a null allele of the female

126 85 germline-specific gene matrimony (mtrm). This allele, referred to as mtrm , was generated through the

86 imprecise excision of a P-element nearly 10 years earlier (Xiang et al. 2007). Presently, the number of B

126 87 chromosomes in this mtrm stock averages 10–12 copies in addition to the A chromosome complement Page 6

88 (Figure 1A and B). Bauerly et al. (2014) demonstrated these B chromosomes have centromeres since

89 they incorporate the Drosophila centromeric histone variant CID and were able to interact with the

90 meiotic spindle. They also analyzed female meiotic prometaphase chromosomes using fluorescent in

91 situ hybridization (FISH) and showed that the B chromosomes carry the AATAT satellite sequence,

92 which is found on the X and Y chromosomes but is primarily on Chromosome 4. Complementation tests

93 and qPCR indicate the B chromosomes do not carry euchromatic material from Chromosome 4, but they

94 do carry large amounts of heterochromatin due to the abundance of histone H3 trimethylation on K9

95 (H3K9m3) and their pronounced effect on position effect variegation (Bauerly et al. 2014).

96 Several questions loom large regarding the origin and composition of the D. melanogaster B

97 chromosomes. It is presently unclear what differences exist, if any, in the size of the B chromosomes as

98 current cytological methods of B chromosome examination do not have the resolution to resolve subtle

99 differences in size. A better grasp of how homo- or heterogenous the size of the B chromosomes are

100 would be informative for understanding the stability of the B chromosomes within the stock.

101 Additionally, the genetic composition of the B chromosomes is poorly understood: it is not clear what, if

102 any, euchromatin the B chromosomes possess, and the known content of its heterochromatin is limited

103 to the AATAT satellite repeat.

104 Here, we report the purification and sequencing of the D. melanogaster B chromosome described

105 in Bauerly et al. (2014). Deep sequencing of this B chromosome reveals no known protein-coding genes

106 but does show that these B chromosomes contain several highly repetitive elements that map to

107 unplaced scaffolds from the current D. melanogaster genome assembly. One of these elements, the

108 previously uncharacterized AAGAT satellite repeat, was used as a FISH target on metaphase

109 chromosome spreads. We find this repeat to be specific to Chromosome 4 as well as enriched twofold on

110 the B chromosomes, suggesting that the B chromosome is likely an isochromosome that arose from

111 Chromosome 4. Surprisingly, FISH analysis of a related stock that carries a different null allele of mtrm

112 revealed the presence of a second, smaller B chromosome. Although present at a much lower copy Page 7

113 number, this newly identified B chromosome variant has the same suite of satellite repeats carried by the

114 original B chromosome, including the AAGAT satellite. Our characterization of these two B

115 chromosomes on a molecular level provides the start of a better understanding of how new

116 chromosomes may arise and become essential fixtures of the genome. Page 8

117 MATERIALS AND METHODS

118 Stocks used in this experiment

119 All stocks used in this report can be found in Table S1. Stocks were maintained on standard cornmeal

120 media supplemented with an additional sprinkle of active dry yeast and kept at 24° and 70% humidity in

121 constant light conditions.

122

123 Sample collection for pulsed-field gel electrophoresis (PFGE)

124 To obtain samples for PFGE, 20 L3 larvae were placed in 3 mL Ringer's solution in a 35-mm glass petri

125 dish. The brain and any attached imaginal discs were dissected from each larva one at a time, and care

126 was taken to ensure the desired tissue was treated as gently as possible to avoid trauma. The tissue was

127 transferred to a 1.5-mL microfuge tube with 500 µL Ringer’s solution kept on ice while the remaining

128 dissections took place. Dissections of 20 larvae took no longer than one hr (typically 30–40 min). Tissue

129 was briefly spun down to the bottom of the tube, the Ringer’s solution was removed, and 30 µL of insert

130 buffer (0.1 M sodium chloride, 30 mM TrisCl pH 8, 50 mM EDTA pH 8, 0.5% Triton X-100, 7.7 mM

131 2-mercaptoethanol) was added. The tissue was homogenized by hand-grinding with a plastic pestle for

132 10 circular strokes. The tube was briefly spun to get the material down, then the homogenization was

133 repeated twice more to ensure complete homogenization. The tube was briefly spun down for a final

134 time and the top-most 25 µL was moved to a new tube in order to minimize the amount of cellular

135 debris included in the plug. This new tube was placed at 42° for 2 min to warm up. Using a cut-off

136 pipette tip (1 mm inner diameter), 75 µL of molten 1% InCert agarose (in 0.125 M EDTA) held at 50°

137 was added to the warmed homogenate and mixed by pipetting gently three times, followed by

138 distribution into a mold for a single plug. The mold was placed at 4° to set and remained there while

139 additional rounds of dissections took place. After dissections were finished, plugs were transferred to Page 9

140 50-mL conical vials (no more than 4 plugs/vial) with 2.5 mL/plug NDS (0.5 M EDTA, 10 mM TrisCl

141 pH 8, 1% sarkosyl, pH 9.5 and sterile-filtered) and 0.1 mg/mL Proteinase K. The tube with the plugs

142 was incubated in a 50° water bath for 12.5 hr and gently swirled once or twice during the incubation.

143 After the Proteinase K treatment, the tubes were placed at 4° until processing, which was usually later

144 the same day (1–2 hr after the end of the incubation).

145 To process the plugs, they were moved from the NDS+Proteinase K solution into new conical

146 tubes with 3 mL/plug 1x TE (pH 7.5) with 0.1 µM PMSF. Tubes were gently agitated on an orbital

147 shaker for 1 hr at room temperature, then the 1x TE + PMSF was replaced with fresh 1x TE + PMSF

148 and the tubes were shaken for another hour at room temperature. This was followed by four washes with

149 1x TE for 20 min each at room temperature, after which the plugs were moved into 1x TAE.

150

151 PFGE setup and parameters

152 To set up the pulsed-field gel rig, 1x TAE was circulated and chilled to 14°. The gel was made with

153 0.8% SeaKem Gold agarose (in 120 mL 1x TAE) that was held at 50° until poured. Sample plugs as

154 well as two plug standards with chromosomes from H. wingei (BioRad) were arranged on the front of a

155 gel comb so they would be embedded in the gel. Approximately 100 mL of the gel was poured and

156 allowed to set for 15 min at 4°. Once the gel was set, the comb was removed, and the empty wells were

157 filled with the remaining unpoured gel and left to set for 5 min. The gel was then placed in the

158 apparatus, followed by the start of the run with the following parameters: 500 s pulse time, 106°

159 included angle, 3 V/cm with linear ramp, 45 hr total run time.

160 Page 10

161 B chromosome DNA Isolation

162 Initially, after the PFGE run (see above), the entire gel was stained in 3x GelRed and imaged to

163 determine the size of the B chromosomes. Once their position was reliably determined, additional PFGE

164 runs were conducted and the B chromosome DNA was extracted as follows. The ladder lanes were cut

165 off and stained for 30 min in 0.5x SYBR Gold in 1x TAE in order to provide a guide for where the B

166 chromosomes were located in the gel. The portion of the gel in the 1.81-Mbp range (where the B

167 chromosomes run) was excised and the remainder of the gel stained to verify the correct part of the gel

168 containing B chromosomes was removed.

169 To extract the B chromosome DNA, the excised gel fragment was placed in dialysis tubing

170 (MWCO = 12K–14K) that was clipped closed at both ends and a second round of electrophoresis was

171 conducted in a normal gel rig. This enables the B chromosome DNA to migrate out of the gel and into

172 the buffer contained within dialysis tubing. Once complete, the buffer containing the B chromosome

173 DNA was pipetted from the tubing and transferred to a 1.5-mL microfuge tube and dried down to

174 concentrate the sample.

175

176 DNA library creation and sequencing

177 The B chromosome DNA sample was sonicated with a Covaris 220 to ~ 600-bp fragments and

178 proceeded directly to library construction using the KAPA High Throughput Library Prep Kit

179 (KK8234), and the Bioo Scientific NEXTFlex DNA Barcodes (NOVA-514104) for barcoding. Quality

180 control was completed using the Agilent 2100 Bioanalyzer and the Invitrogen Qubit 2.0 Fluorometer.

181 The sample was sequenced on the Mi-Seq via a Mi-Seq 150-bp paired-end run. This run used MiSeq

182 Reporter Control Software (v2.6.2.3) and RTA version MCS 2.6.1.1.

183 Page 11

184 Genome alignment, assembly, and analysis

185 Reads were aligned to dm6 using bwa (Li and Durbin 2009). Unmapped reads were identified by

186 ‘samtools view -f4’ and unmapped read pairs were isolated and trimmed with sickle

187 (https://github.com/najoshi/sickle) and scythe (https://github.com/ucdavis-bioinformatics/scythe),

188 removing bases with quality scores < 30 and excluding any read < 40 bp after trimming. SOAPdeNovo2

189 was used to assemble unmapped reads using a kmer size of 41 and average insert size of 300 (Luo et al.

190 2012). Assembly generated 347,712 contigs, 75 of which were larger than 500 bp with a maximum

191 contig size of 1,793 bp. To determine whether the 75 contigs larger than 500 bp contained protein-

192 coding sequence, BLASTx (Altschul et al. 1990) was used to search RefSeq (update date 2018-04-02,

193 containing 103,725,652 sequences) (O’Leary et al. 2016). We used k-Seek (Wei et al. 2014) to identify

194 kmers present on the isolated B chromosomes, adding a third step that we call k.compile which

195 summarizes the data generated by the original implementation of the software.

196

197 Chromosome squashes

198 For DNA staining and FISH: For chromosome spreads from larval brain tissue, a protocol adapted

199 from Larracuente and Ferree (2015) was used. Before beginning, 2 mL of fixative solution (45% acetic

200 acid, 2.5% paraformaldehyde) was made fresh. A single larva was removed and placed in a fresh 50-µL

201 drop of 0.7% sodium chloride. The brain and associated imaginal discs were extracted in the same

202 fashion as for the PFGE (see above). The brain was then moved to a fresh 50-µL drop of 0.5% sodium

203 citrate for hypotonic treatment for 5 min, followed by a transfer to the fixative solution for 4 min. After

204 fixation, the brain was transferred to a 3-µL drop of 45% acetic acid on an 18-mm x 18-mm No. 1.5

205 siliconized coverslip. A microscope slide was inverted onto the coverslip with the sample and pressed

206 gently to spread the liquid to the edges of the coverslip. The slide+coverslip was squashed for 2 min

207 using a hand clamp (Milwaukee Tools, 48-22-3002), then immediately placed into liquid nitrogen for at Page 12

208 least 5 min. Immediately after removal, the coverslip was popped off the slide using a razor blade. The

209 slide was dehydrated by placing it in 70% ethanol for at least 10 min at -20°, then transferred to 100%

210 ethanol at -20° for at least 10 min. Slides were removed and allowed to completely air-dry, then stored

211 in a corked slide box kept at room temperature.

212 For chromosome spreads from ovary tissue, the protocol used for larval brains described above

213 was used with the following changes. Ovaries were collected from mated adult females by anesthetizing

214 them with carbon dioxide until incapacitated (~5–10 seconds). A single female was then moved to a

215 fresh 50-µL drop of 0.7% sodium chloride and whole ovaries were removed. The tips were separated

216 from the later stages such that anything larger than stage 8 was discarded. The ovary tips were then

217 hypotonically treated and fixed as above for larval brains. Before squashing, the tips were gently teased

218 apart on the siliconized coverslip in 3 µL 45% acetic acid in order to spread out the tissue. Squashing,

219 freezing, dehydration, drying, and storage were all carried out as described above for larval brains.

220 For immunofluorescence (IF): To preserve protein-DNA interactions in metaphase

221 chromosome spreads from larval brain tissue, a protocol similar to the acid-free squash technique in

222 Johansen et al. (2009) was used. Before beginning, 2 mL of Brower’s fixative solution (0.15 M PIPES, 3

223 mM magnesium sulfate, 1.5 mM EGTA, 1.5% Nonidet P-40 substitute, 2% paraformaldehyde) was

224 made fresh. Larval brains were removed in 0.7% sodium chloride and incubated in 2 mL 50 µM

225 colchicine in 0.7% sodium chloride at 25° for 1 hr. (The colchicine arrests cells in metaphase, resulting

226 in super-condensed chromosomes that are able to hold their shape better after Brower’s fixative.) After

227 the incubation, the brain was transferred to a fresh 50-µL drop of 0.5% sodium citrate for hypotonic

228 treatment for 5 min, followed by a transfer to the fixative solution for 2 min. The brain was then moved

229 to a 50-µL drop of 50% glycerol for 5 min, then to 5-µL drop of 50% glycerol on an 18-mm x 18-mm

230 No. 1.5 siliconized coverslip. The brain tissue was gently dissociated using a bent needle probe, then a

231 microscope slide was inverted onto the coverslip and direct pressure was applied to spread the tissue. Page 13

232 The slide+coverslip was squashed, frozen, and dehydrated as described above. Slides were allowed to

233 air-dry for 10 min, followed by immediate processing for IF.

234

235 Sample slide processing

236 For DNA staining and FISH: For simple DNA visualization, 5 µL of Vectashield with DAPI mounting

237 media was applied to a clean 22-mm x 22-mm No. 1.5 glass coverslip that was placed on the sample

238 slide,then sealed to the slide with clear nail polish.

239 For FISH on the chromosome spreads, a protocol adapted from Larracuente and Ferree (2015)

240 was used. For each slide, 21 µL of the FISH solution (50% formamide, 10% dextran sulfate, 2x SSC,

241 100 ng fluorescently-labeled probe (see Table S2) was applied directly to the dried slide and a clean 22-

242 mm x 22-mm No. 1.5 glass coverslip was placed on top. Gently, any large bubbles that in contact with

243 the tissue were nudged to the edges. The slide was heated to 95° on a heat block for 5 min in darkness,

244 then transferred to a container lined with damp paper towels (to maintain humidity and prevent the

245 samples from drying out) and placed at 30° overnight (16–24 hr). After the incubation, slides were

246 washed three times in 0.1x SSC for at least 15 min each. Slides were blown dry, then mounted by

247 applying 5 µL Vectashield with DAPI to a clean 22-mm x 22-mm No. 1.5 glass coverslip that was

248 placed on the sample slide, then sealed to the slide with clear nail polish.

249 For IF: Dried slides were placed in PBSTX (1x PBS with 0.1% Triton X-100) and washed three

250 times for 5 min each. Slides were then blocked with 5% dry, nonfat milk in PBSTX for 20 min at room

251 temperature. Excess block was wiped away from the perimeter of the sample area and a primary

252 antibody solution (antibody in 30 µL block solution; 1:500 rat anti-CID from a test bleed generated for

253 the Hawley lab, 1:100 rabbit anti-HOAP serum, a gift from Y. Rong) was applied, and a 22-mm x 22-

254 mm No. 1.5 glass coverslip was placed on top. The slides were placed in a container lined with damp

255 paper towels and placed at 4° overnight (16–24 hr). After the incubation, the slides were washed with Page 14

256 PBSTX three times for 5 min each. Excess PBSTX was wiped away from the perimeter of the sample

257 area and a secondary antibody solution (antibody in 30 µL block solution; 1:500 anti-rat from

258 Invitrogen, 1:500 anti-rabbit from Invitrogen) was applied, and a 22-mm x 22-mm No. 1.5 glass

259 coverslip was placed on top. The slides were placed in a container lined with damp paper towels and

260 placed in a dark drawer at room temperature for 45 min. After the incubation, the slides were washed

261 with PBSTX three times for 5 min each. Excess PBSTX was wiped away from the perimeter of the

262 sample area and 3 µL Vectashield with DAPI was applied to a clean 22-mm x 22-mm No. 1.5 glass

263 coverslip that was placed on the sample slide, then sealed to the slide with clear nail polish.

264

265 Microscopy and image processing

266 Images were acquired with a DeltaVision microscopy system (GE Healthcare, Piscataway, NY)

267 consisting of a 1x70 inverted microscope with a high-resolution CCD camera. All imaging used a 100x

268 objective and 1.6x auxillary magnification. Stacks of 15 z images with a thickness of 0.2 µm were taken

269 for each metaphase using 100% transmission and standard DAPI, FITC, and TRITC filters. Images were

270 deconvolved using SoftWoRx v. 6.1.3 or later (Applied Precision/GE Healthcare) following Applied

271 Precision protocols. Stacks of deconvolved images were combined in a z projection showing maximum

272 intensity, cropped to the region of interest, recolored, and adjusted for brightness and contrast in

273 FIJI/ImageJ.

274 For direct comparisons of AAGAT FISH signal, samples were collected, prepped, and imaged

275 concurrently. Settings for excitation/emission/exposure were identical for all four samples. Stacks of

276 deconvolved images were combined in a z projection showing maximum intensity, cropped to the region

277 of interest, and recolored in FIJI/ImageJ. Only the DAPI channel was adjusted for brightness and

278 contrast; no adjustments were made to the FITC and TRITC channels, thus the display range of each

279 image in those channels represents the full, original pixel value range. Page 15

280

281 Intensity measurements of DAPI and FISH probes

282 After imaging (as described above), deconvolved image stacks were opened in FIJI/ImageJ and made

283 into a z projection by summing the intensity of the slices in the stack (instead of maximum intensity

284 projections) in order to preserve the total intensity. A circular region of a defined size was drawn around

285 chromosomes (either B1, B2, or Chromosome 4) that were well separated from other chromosomes. An

286 equal number of regions were placed near the circled chromosomes in order to obtain background

287 intensity values.

288 Integrated density measurements from both the DAPI and FISH channels within each region

289 were taken, and measurement values from like chromosomes were averaged for each metaphase

290 individually; for example, each of the measurements for the B1 chromosomes in a single metaphase

291 were averaged, as well as the values for the background measurements within the same single

292 metaphase. When there was only one of a chromosome, the measured value was used. To adjust for the

293 background, the average of the background measurements was subtracted from the average (or single

294 value) of the chromosomes. These adjusted values were then divided by one another in order to get the

295 appropriate statistic.

296

297 Data availability and Supplemental Material

298 Illumina sequencing reads generated in this report are available at the National Center for Biotechnology

299 Information (https://www.ncbi.nlm.nih.gov/) under project PRJNA484270. Scripts used in the analysis

300 of this project can be found at https://github.com/danrdanny/B-chromosome/. Original data underlying

301 this manuscript can be accessed from the Stowers Original Data Repository at

302 http://www.stowers.org/research/publications/LIBPB-1334. Supplemental files available at FigShare. Page 16

303 Supplemental File S1 includes all supplemental figures (Figures S1–4) and tables (Tables S1–4).

304 Supplemental File S2 are blast results. All stocks and reagents available upon request. Page 17

305 RESULTS

306 Molecular isolation of D. melanogaster B chromosomes

307 The B chromosomes reported in Bauerly et al. (2014) were first identified in a stock carrying a null

126 308 allele of matrimony (mtrm ) generated by the Hawley laboratory around 2004. Matrimony is a Polo-

309 kinase inhibitor that is specific to female meiosis (Xiang et al. 2007). The level of Mtrm, relative to Polo

310 kinase, is important during oogenesis: half the wild-type amount of Mtrm results in the missegregation

311 of noncrossover chromosomes during meiosis I, and the complete absence of Mtrm leads to meiotic

126 312 catastrophe and sterility (Bonner et al. 2013). Individual flies within this mtrm stock possess an

313 average of 10–12 B chromosomes as assessed by examining metaphase chromosomes from brains of

314 third-instar larvae (Figures 1A and B).

315 Due to the increased potential for an aberrant female meiosis in this stock, we wondered if the B

316 chromosomes were the product of a chromosomal fragmentation event (or events) that may have

317 occurred following a mtrm-induced premature entry into meiotic anaphase I (Bonner et al. 2013). If the

318 present B chromosomes are copies of an original single, stable chromosome fragment, then we expect

319 all B chromosomes to be the same size, whereas if each B chromosome was a product of a different

320 fragmentation event, each one would likely be a different size. Because the current cytological method

321 of B chromosome analysis is not sensitive enough to detect small size differences between B

322 chromosomes, we looked for a method that would allow a more precise size measurement of the B

323 chromosomes. Pulsed-field gel electrophoresis (PFGE), a technique routinely used to resolve large DNA

324 fragments in the megabase range, enabled us to measure the size(s) of intact B chromosomes. Brains and

126 325 imaginal discs from male and female third-instar larvae in the mtrm stock were disrupted and

326 embedded in agarose, with all proceeding treatments occurring in-gel to prevent shearing of whole

327 chromosomes (see Materials and Methods). Our PFGE produced excellent resolution within the

328 approximate size range of the B chromosome, which appeared as a single, distinct band around 1.81 Page 18

329 Mbp (Figure 1C). We did not see a smear on the gel that would represent a multitude of B chromosomes

330 sizes, suggesting the B chromosomes are likely stable and not continually fusing and fragmenting. We

331 also did not observe additional discrete bands in addition to the band at 1.81 Mbp, leading us to

332 conclude that the B chromosomes in this stock are of a very similar size.

333

334 Sequence analysis of isolated B chromosomes confirms their heterochromatic composition

335 Isolation of the D. melanogaster B chromosomes by PFGE gave us an opportunity to study them in the

336 absence of the standard chromosome complement. After PFGE, the portion of the gel encompassing the

337 B chromosomes was removed and the DNA extracted, purified, and sequenced via paired-end Illumina

338 sequencing (see Materials and Methods). We generated 42,964,020 150-bp paired-end reads (21,482,010

339 read pairs) and aligned them to release 6 (dm6) of the D. melanogaster genome. A total of 30%

340 (13,031,938) of all reads aligned in pairs to unique locations in the reference genome, while 33%

341 (14,212,886) of reads mapped to either non-unique locations or were pairs that did not align properly.

342 The remainder of the reads (37%, 15,719,196) did not map to known genomic locations.

nd rd 343 Among the 13,031,938 uniquely mapping reads, 84% aligned to the X, 2 , 3 , or Y

344 chromosomes with an average depth of coverage of 12x (Table S3). We anticipated this low, evenly

345 distributed depth of coverage across the genome since our PFGE purification would likely include a

346 random assortment of genomic fragments that were in the size range of the B chromosome. We then

347 masked repetitive sequence using RepeatMasker and plotted depth of coverage to identify regions that

348 exceeded this background (Smit et al. 2015). We found no focal increases in coverage along the

349 euchromatic regions of the genome, confirming that the B chromosomes do not carry any euchromatic

350 genomic sequence (Figure S1). As for Chromosome 4, 5% of the uniquely mapping reads aligned with

351 an average depth of coverage of 76x. Masking repetitive sequence reduced this to approximately 10x,

352 similar to the genome-wide background. This reduction in coverage demonstrates that many reads that Page 19

353 uniquely mapped to Chromosome 4 are in repetitive or low-complexity areas. This observed lack of

354 euchromatic sequence in our isolated B chromosome sample confirms previous work that suggested the

355 B chromosomes are highly heterochromatic (Bauerly et al. 2014).

356 The remaining 11% of uniquely mapped reads aligned to small reference scaffolds whose exact

357 position in the genome is unknown. These unplaced scaffolds are generally short, highly repetitive, or

358 low complexity fragments, suggesting they are heterochromatic sequences. Of 1,870 scaffolds present in

359 the dm6 assembly, we found 133 that contained at least one region with 100x depth of coverage or

360 greater in uniquely mapping reads, suggesting these scaffolds contain sequence present on the B

361 chromosomes. In total, 114,673 bp of total sequence across these 133 contigs had at least 100x depth of

362 coverage (Table S4). As an example, the 51-kb scaffold chrUn_DS483562v1 had at least 100x depth of

363 coverage in uniquely aligning reads over more than 80% of its sequence, suggesting that it contains a

364 large, but not necessarily contiguous, amount of DNA present on the B chromosomes.

365 We also wondered if the B chromosomes carried any novel sequence—euchromatic or

366 heterochromatic—not present in the current release of the Drosophila genome. To determine if any of

367 the 15,719,196 unmapped reads were from unique, non-repetitive sequences not represented within the

368 D. melanogaster genome assembly, we performed a de novo assembly using SOAPdenovo2 (Luo et al.

369 2012) using a kmer size of 41 (see Materials and Methods). This resulted in the assembly of 75 contigs

370 greater than or equal to 500 bp with an average assembled length of 744 bp (max 1,793 bp). We then

371 searched RefSeq using BLASTx to determine if these contigs contained any complete genic sequences

372 or other known sequence and found that 39 of 131 contigs contained at least one significant hit in the

373 RefSeq database. Of all the proteins identified using our method, none were complete copies and the

374 majority were viral, suggesting that several of these contigs contained fragments of transposable

375 elements (File S2). Taken together, this sequence analysis suggests that the B chromosomes we

376 sequenced are derived from the D. melanogaster genome and do not contain any previously known or

377 novel euchromatic sequence. Page 20

378

379 FISH analysis indicates Chromosome 4 is the source of the B chromosomes

380 Previous analysis of female meiotic prometaphase chromosomes using FISH revealed the B

381 chromosomes carry the simple nucleotide repeat AATAT. Though the B chromosomes are able to move

382 independently during prometaphase, the remainder of the chromosomes form a mass that makes it

383 difficult to discern where a repetitive sequence resides on an individual chromosome. Additionally,

384 since the Y chromosome is absent in female prometaphases, we cannot perform extensive FISH analyses

385 to rule it out as a potential source of the B chromosomes. To circumvent these obstacles, we performed

386 FISH using the same AATAT probe on metaphase chromosomes from male third-instar larval brains. We

387 were able to verify the previously described presence of this repeat on the B chromosomes, as well on

388 Chromosome 4, the heterochromatic tip of the X, and several locations on the Y chromosome (Lohe et

389 al. 1993; Bauerly et al. 2014) (Figure 2A).

390 Since the B chromosome may have arisen from any of these chromosomes, we wondered if our

391 B chromosome sequence contained short, tandemly repeating kmers that could be used as FISH targets.

392 We used k-Seek (Wei et al. 2014) to generate a list of kmers present in the B chromosome raw

393 sequencing data (Table 1). Among the kmers with over a million counts, we recovered the AATAT

394 repetitive sequence, confirming our previous FISH results. Additionally, we found the AAGAG

395 repetitive sequence was enriched on the B chromosomes as well as Chromosome 4 (Figure S2A),

396 however this repeat can also be found on each of the other chromosomes as it is one of the most

397 abundant 5mer repeats in the genome (Lohe et al. 1993; Wei et al. 2014). Based on these genome-wide

398 studies, we examined other abundant repetitive sequences that were not prevalent in our kmer analysis

399 and found they do not appear to be present on the B chromosomes (Figure S2B-F). The most prevalent

400 repetitive sequence identified in our kmer analysis was the 5mer AAGAT, which has appeared recently in

401 two genomic sequencing datasets but has yet to be fully characterized (Wei et al. 2014; Talbert et al. Page 21

402 2018). Using a FISH probe for the AAGAT repeat, we found it to be highly enriched on the B

403 chromosomes as well as at the tip of Chromosome 4 (Figures 2B and 2C). As Chromosome 4 is the only

404 chromosome that shares all three of the most abundant repeats found on the B chromosome, it is likely

405 to be the source of the B chromosomes.

406

407 B chromosomes have functional telomeres that recruit the capping protein HOAP

408 Due to its smaller size and its similarity in composition, the initial B chromosome may have been a

409 fragment that broke away from Chromosome 4. A double-strand break on Chromosome 4 would

410 produce two fragments, each with a single telomere and a single broken end. Only one of these

411 fragments, however, would possess the centromere and be able to be segregated during cell division. For

412 the centric fragment to be adequately maintained, its broken end would need to be mitigated, such as

413 through the formation of a ring chromosome that would negate a need for telomeres altogether (Titen

414 and Golic 2008). Alternatively, the broken end may become a functional neo-telomere that can recruit

415 telomeric capping proteins, which has previously been observed in Drosophila (Gao et al. 2010;

416 Kurzhals et al. 2017).

417 To distinguish between these possibilities, we wanted to determine if the B chromosomes have

418 functional telomeres by looking for the presence of the telomere-specific capping protein HOAP

419 (encoded by cav) on chromosomes in metaphase spreads (Cenci et al. 2003). Though the protocol used

420 for FISH analyses produces excellent chromosome morphology, acetic acid in the fixative solution is

421 known to remove proteins associated with DNA, such as histones (Dick and Johns 1968; Pimpinelli et

422 al. 2011). To avoid this, we switched to a protocol similar to the acid-free squash technique in Johansen

423 et al. (2009). Using antibodies that recognize the Drosophila centromeric histone CID or the telomere

424 capping protein HOAP, we conducted immunofluorescence (IF) on acid-free metaphase chromosome

425 squashes. The B chromosomes are able to incorporate CID (Figure 3), confirming previous results Page 22

426 (Bauerly et al. 2014). We observed HOAP localization at the ends of all the chromosomes, including the

427 B chromosomes, leading us to conclude they are linear chromosomes with functional telomeres (Figure

428 3).

429

430 Location and abundance of the AAGAT satellite repeat on the B chromosome is consistent with the

431 structure of an isochromosome

432 During our FISH analysis, we noticed that the number of fluorescent foci observed on Chromosome 4

433 with the AAGAT probe was never more than two (which we interpret as one focus per chromatid),

434 whereas within the same metaphase, the B chromosomes could have as many as four foci. This

435 observation is not dependent on degree of chromosome condensation as we do not treat the tissue with a

436 cell cycle inhibitor prior to performing FISH (Figure 4A-C). If the additional FISH foci are due to the

437 AAGAT satellite repeat block being split on the B chromosomes, then the intensity of the FISH signal

438 should be similar to that of Chromosome 4. However, if there is more of the AAGAT satellite repeat on

439 the B chromosomes, then their FISH signal would be greater than that of Chromosome 4. We therefore

440 measured the intensity of the AAGAT FISH signal on individual B chromosomes and compared it to the

441 signal on individual Chromosome 4s within the same metaphase (see Materials and Methods). After

442 analyzing 29 metaphases, we found a twofold difference in the AAGAT FISH probe signal intensity

443 when comparing the B chromosomes to Chromosome 4 (Figure 4D).

444 This intensity is twice what we would expect if the B chromosome was derived from a single

445 chromosome fragment, leading us to reassess how the D. melanogaster B chromosome may have arisen

446 from Chromosome 4. We considered the reported structures of B chromosome in other species, such as

447 the B chromosome in the characid fish Astyanax scabripinnis. This B chromosome appears to be an

448 isochromosome that is thought to have formed after a centromeric misdivision event (Mestriner et al.

449 2000). Such an event separates the left arms from the right arms of a chromatid pair due to a break at the Page 23

450 centromere; if the two identical arms fuse together, they simultaneously reconstitute the centromere and

451 heal the broken chromosomes ends, resulting in an isochromosome. Focusing on the apparent

452 duplication of the AAGAT satellite repeat on the B chromosomes as compared to Chromosome 4, we

453 propose the D. melanogaster B chromosome is an isochromosome that formed after a centromeric

454 misdivision event on Chromosome 4 (see Discussion).

455

456 Identification of a second B chromosome variant

126 457 Recently, we began examining the progenitor of the mtrm stock that is retained in our laboratory. This

458 stock carries a different null allele of mtrm caused by a P-element insertion upstream of the open reading

KG08051 459 frame, herein designated mtrm (Harris et al. 2003; Xiang et al. 2007). In a few of our samples

460 from this stock, the metaphase chromosome spreads included a small DAPI-staining fragment that was

461 reminiscent of a B chromosome. To determine if this DNA fragment had a similar repeat composition as

462 the B chromosome, we performed FISH on metaphase spreads using probes for repeat sequences that are

463 enriched on the B chromosome, revealing that this small DNA fragment carries each of them (Figure

464 5A, Figure S3). Since this DNA fragment was not present in every sample we examined, we speculated

465 it may lack essential chromosome components such as a centromere and/or telomeres, thus affecting its

466 ability to be efficiently transmitted to progeny. Therefore, we performed IF on metaphase spreads from

KG08051 467 the mtrm stock and found that this element carries both the CID and HOAP antigens, suggesting

468 that it possesses both centromeres and telomeres, respectively (Figure 5B). Taken together, we believe

469 this DNA fragment is the second B chromosome variant identified in D. melanogaster and will herein

470 refer to it as the “B2” chromosome to distinguish it from the original B chromosome (now referred to as

471 “B1”).

472 Though the B1 and B2 chromosomes have a similar composition, there are some apparent

473 cytological differences that exist between them. FISH experiments with the AAGAT probe showed that Page 24

474 the B2 chromosome has fewer FISH foci—as well as less overall DAPI intensity—than does the B1

475 chromosome. To directly compare the two B chromosomes to each other as well as to the same

476 Chromosome 4, a male carrying the B1 chromosome was crossed to a female carrying the B2

477 chromosome. Progeny from this cross would possess both B chromosomes, enabling their direct

478 comparison since they are within the same metaphase (Figure 6A). (We should note that the quantitative

479 comparisons between the B1 chromosome and Chromosome 4 in Figure 4D were performed on the

480 same metaphases, therefore the Chromosome 4 used for comparison here is identical.) Comparing the

481 AAGAT FISH signal intensity on the B2 chromosome to Chromosome 4 demonstrates that they have an

482 equal amount of FISH signal (Figure 6B, see Materials and Methods). This result is consistent with our

483 observation that the B2 chromosome has fewer FISH foci than the B1 chromosome (Figure 6C). Since

484 both B chromosome variants carry a similar suite of satellite repeats and are likely similar in their DNA

485 composition, we used their DAPI staining intensity to relatively compare their DNA content. We found

486 that the B2 chromosome is 65% as bright as the B1 chromosome, suggesting the B2 chromosome has

487 less DNA content overall and is likely smaller in total length than the B1 chromosome (Figure 6D, see

488 Materials and Methods). Due to these differences, we conclude that the B1 and B2 chromosomes are

489 distinct and represent two variants of B chromosomes in D. melanogaster.

490

KG08051 491 Other stocks carrying the mtrm allele do not have B chromosomes and lack the AAGAT

492 repeat

KG08051 493 The mtrm stock carrying the B2 chromosome was originally a gift from the Bellen lab and has

494 been in our laboratory collection since 2003 (Harris et al. 2003). Copies of this stock carrying the same

KG08051 495 mtrm null allele also exist in our institute stock collection as well as the Bloomington Drosophila

SIMR1 SIMR2 BDSC 496 Stock Center (herein referred to as mtrm , mtrm , and mtrm , respectively; see Table S1). To

497 determine if any of these three lines also carry a B chromosome and at what frequency, we examined Page 25

498 them cytologically and looked for small DAPI-staining fragments. Although previous chromosome

499 spreads described here were made from larval neuroblast tissue, for this analysis we chose to use ovarian

500 tissue from adult females in order to quickly assess how many, if any, breeding adults carried B

KG08051 501 chromosomes within each of the stocks. We began by examining the mtrm line that we know

502 carries the B2 chromosome in order to determine its frequency within the stock. We found that nine out

503 of 25 adult females carried one copy of the B2 chromosome and two females carried two copies of the

504 B2 chromosome, indicating its frequency within the stock is very low (Figure S4A). We then examined

505 females from the other three stocks using the same technique and included the AATAT FISH probe in

506 our chromosome spread processing since it is enriched on both the B1 and B2 chromosomes.

507 Interestingly, after examining at least 30 females from each stock, we did not detect a supernumerary

SIMR1 SIMR2 508 DAPI-staining fragment (Figure S4B-E). We therefore conclude that the mtrm , mtrm , and

BDSC 509 mtrm lines do not carry B chromosomes.

SIMR1 SIMR2 BDSC 510 During the initial examination of the mtrm , mtrm , and mtrm lines, we used the

511 AAGAT FISH probe in our chromosome spread processing. We ultimately switched to the AATAT probe

512 for our analyses, however, since we were unable to detect AAGAT FISH signal in the other three lines.

513 This result was puzzling for two reasons. First, though these three stocks were acquired from separate

KG08051 514 collections, they are genotypic copies of our laboratory mtrm stock that clearly carries the AAGAT

515 satellite repeat (Figure 7A) and therefore we expected them to be similar. Second, since the AAGAT

516 satellite repeat was abundant in two separate genomic studies (Wei et al. 2014; Talbert et al. 2018), we

517 were surprised to not detect any blocks of AAGAT satellite repeat in these three stocks. To validate our

518 original observation, we repeated the FISH experiment on neuroblast chromosome spreads from males

519 (in order to also assay the Y chromosome) using both the AAGAT and AATAT probes simultaneously.

520 Consistent with our initial results, we observed FISH signal with the AATAT probe but not the AAGAT

521 probe in these three stocks (Figure 7B-D). This result confirms that the lack of AAGAT signal we Page 26

522 observe is due to polymorphism in the amount of AAGAT satellite repeat carried by the variants of

523 Chromosome 4 in these stocks. Page 27

524 DISCUSSION

525 In the century since the first description of a supernumerary chromosome, the existence of B

526 chromosomes has been well documented in several species spanning a diverse range of taxa. Despite

527 their widespread prevalence, only a handful of systems have possessed tools to enable an in-depth

528 molecular and genomic analysis of their B chromosome(s) (López-León et al. 1994; Han et al. 2007;

529 Banaei-Moghaddam et al. 2012; Kour et al. 2014; Ellis et al. 2015; Huang et al. 2016; Aldrich et al.

530 2017; Ruiz-Ruano et al. 2018). The appearance of B chromosomes in an established model animal

531 system such as D. melanogaster enables us to use established molecular tools to study the anatomy of a

532 newly formed chromosome and quickly ascertain its origin. By isolating and deep-sequencing the B1

533 chromosome, we were able to determine that it originated from the D. melanogaster genome, lacks

534 known euchromatic material, and is comprised of several heterochromatic elements.

535 One of the heterochromatic elements we identified is the AAGAT satellite repeat, which has been

536 recently reported but was not characterized (Wei et al. 2014; Talbert et al. 2018). The AAGAT repeat

537 was found to be the fourth-most abundant 5mer repetitive sequence in a whole-genome analysis of

538 several Drosophila lines (Wei et al. 2014), and a separate study concluded it was one of the major D.

539 melanogaster centromere components due to its enrichment in CENP-A ChIP experiments (Talbert et al.

540 2018). Though Talbert et al. (2018) suspected it as a candidate for the centromeric satellite repeat on the

541 Y chromosome, we did not detect AAGAT sequence on the Y chromosome via FISH in any of our stocks,

542 including the stocks that lacked the AAGAT satellite repeat on Chromosome 4 (Figure 7). Rather, we

543 have shown here that this repeat appears to be uniquely localized to Chromosome 4 and to both B

544 chromosomes. Since the AAGAT repeat is associated with the centromeric histone CENP-A (Talbert et

545 al. 2018), and our chromosome cytology demonstrates the presence of AAGAT on Chromosome 4 and

546 the B chromosomes, we speculate the AAGAT satellite sequence is associated with the centromere on Page 28

547 Chromosome 4. Additional studies are necessary to definitively conclude if this satellite is indeed a

548 functional part of the Chromosome 4 centromere.

549 Based on the AAGAT satellite repeat cytology, the structure of the B1 chromosome is most

550 consistent with that of an isochromosome formed from Chromosome 4 (Figure 8). It is likely comprised

551 of the shorter, left arm of Chromosome 4 since the right arm carries euchromatic material, which has

552 been shown both previously (Bauerly et al. 2014) and in this study to be absent from the B

553 chromosomes. One well-documented mechanism of isochromosome formation is through a centromeric

554 misdivision event (Rhoades 1933; Upcott 1937; Darlington 1939; Carlson 1970; Kaszás and Birchler

555 1996). The resulting chromosome fragments can fuse to become an isochromosome, thus eliminating the

556 broken ends and potentially reconstituting a full centromere. It has been demonstrated in wheat that the

557 formation of isochromosomes from univalent (unpaired) chromosomes through a centromeric

558 misdivision event is quite frequent during meiotic divisions (Sears 1952; Steinitz-Sears 1966; Friebe et

559 al. 2005; Lukaszewski 2010). Such an event may have the opportunity to occur more often in a mtrm

560 mutant background due to an increased frequency of Chromosome 4 aneuploidy within the stock, a

561 result of its frequent missegregation during female meiosis (Harris et al. 2003; Xiang et al. 2007;

562 Bonner et al. 2013). Females carrying an extra copy of Chromosome 4 would produce a univalent

563 during meiosis, which may be more susceptible to centromeric misdivision. Additionally, B

564 chromosomes carried by other species have been shown to be isochromosomes that arose from one of

565 the essential A chromosomes (John and Hewitt 1965; Vicente et al. 1996; Mestriner et al. 2000; Dhar et

566 al. 2002; Poletto et al. 2010). Indeed, as suggested by Battaglia (1964), misdivision and isochromosome

567 formation is likely a recurring mechanism for B chromosome creation.

568 We also identified a second, smaller B chromosome variant (B2) in a separate laboratory stock

569 carrying a different null allele of mtrm. Though the B2 chromosome carries the same suite of repeats as

570 the B1 chromosome, has both telomeric and centromeric components, and was found in a similar mutant

571 background, the copy number and frequency of the B2 chromosome in its endogenous stock Page 29 KG08051 572 (mtrm ) is substantially lower than is characteristic of the B1 chromosome copy number in its

126 573 endogenous stock (mtrm ). This observation may be attributed to the variation in genetic backgrounds

574 (see Table S1) or the difference in the size of the two B chromosomes. Periodic testing of this stock will

575 enable us to use established methods to determine if the prevalence of this new B chromosome is

576 fluctuating and, if so, at what rate.

577 Though the B1 chromosome structure is consistent with an isochromosome, we are unsure what

578 the structure of the B2 chromosome may be. Since the DAPI intensity of the B2 chromosome is 65%

579 that of the B1 chromosome (Figure 6D), we favor a model where the B2 chromosome arose after a

580 centromeric misdivision event that encompassed a larger portion of the centromeric region, followed by

581 the acquisition of a telomere instead of the joining of the broken fragments. We are aware, however, that

582 due to the shared histories of the stocks each B chromosome was found in, we cannot establish whether

583 each B chromosome arose independently, or if one was formed from another. We did not attempt to

584 isolate the B2 chromosome after PFGE as we anticipate its size and scarcity will be a limiting factor for

585 the amount of material we can recover. However, newer sequencing technologies that produce

586 extremely long reads may enable us to circumvent this limitation and determine the sequence of the

587 entire B2 chromosome without the need to isolate it first. Variations such as single-nucleotide

588 polymorphisms, small insertions and deletions, or transposable element positions that differ between the

589 B chromosomes and Chromosome 4 may be used to distinguish the order of origin of the B

590 chromosomes and help determine what, if any, constraints exist on B chromosome evolution.

591 Additionally, long-read sequencing of the satellite repeat regions will produce a more accurate analysis

592 of their composition, allowing us to track their changes over time and in various genetic backgrounds.

593 Our results strongly indicate the B chromosomes originated from Chromosome 4, however we

594 presently are unaware as to how often they arise and how long they can be maintained. To this end, we

126 595 are curious what may make the original mtrm stock (with B1 chromosomes) and the single

KG08051 596 mtrm stock (with B2 chromosomes) unique. Though all five stocks we examined carry a null Page 30

597 allele of mtrm (Table S1), only the stocks with B chromosomes also have the AAGAT satellite repeat on

598 Chromosome 4 (as assessed by FISH on metaphase chromosome spreads, Figure 7). We are intrigued as

599 to what role, if any, this repetitive sequence plays in the formation and/or maintenance of a B

600 chromosome.

601 Ultimately, a deeper understanding of the structure and formation of the D. melanogaster B

602 chromosomes may provide insight into the etiology of similar chromosomes that can be detrimental to

603 human health. Small supernumerary marker chromosomes (sSMCs) in humans have many properties

604 that are similar to B chromosomes (Fuster et al. 2004; Liehr et al. 2008). These sSMCs arise from the A

605 chromosomes and are present in 1 out of ~2,300 newborns (Liehr and Weise 2007). Many are connected

606 to specific syndromes, such as Emanuel, cat eye, and Pallister-Killian syndromes, or are associated with

607 intellectual disabilities and infertility (Liehr 2012; Armanet et al. 2015; Jafari-Ghahfarokhi et al. 2015;

608 Wang et al. 2015). Improved molecular characterization is beginning to reveal more about the genomic

609 organization of supernumerary chromosomes (Breman et al. 2011; Sun et al. 2017). A subset of sSMCs

610 are isochromosomes derived from two of the short arms of a single chromosome after an apparent

611 centromeric misdivision event (de la Chapelle 1982; Bugge et al. 2004; Jafari-Ghahfarokhi et al. 2015).

612 Isochromosomes are also observed in various cancers, most notably testicular germ cell tumors where

613 isochromosome 12p is a well-established hallmark and is believed to be a triggering event for invasive

614 growth (Litchfield et al. 2016; Mitelman et al. 2018). Continued investigation of the mechanisms behind

615 supernumerary chromosome formation, retention, and transmission using D. melanogaster B

616 chromosomes as a model system will thus be of interest to a wide range of research areas. Page 31

617 ACKNOWLEDGEMENTS

618 We would like to thank the Hawley Laboratory for their support and helpful comments on the

619 manuscript. We also thank Rhonda Egidy, Anoja Perera, and the Molecular Biology core at SIMR for

620 expert help with DNA sequencing. The rabbit anti-HOAP serum was a generous gift from Yikang S.

621 Rong at Sun Yat-sen University in China, and we thank him for sharing it with us. RSH is supported by

622 the Stowers Institute for Medical Research and is an American Cancer Society Research Professor. Page 32

623 REFERENCES

624 Aldrich J. C., A. Leibholz, M. S. Cheema, J. Ausiό, and P. M. Ferree, 2017 A “selfish” B chromosome

625 induces genome elimination by disrupting the histone code in the jewel wasp Nasonia vitripennis.

626 Sci Rep 7: 42551. https://doi.org/10.1038/srep42551

627 Altschul S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman, 1990 Basic local alignment search

628 tool. J Mol Biol 215: 403–10. https://doi.org/10.1016/S0022-2836(05)80360-2

629 Araújo S. M. S. R., S. G. Pompolo, F. Perfectti, and J. P. M. Camacho, 2002 Integration of a B

630 chromosome into the A genome of a wasp, revisited. Proceedings Biol Sci 269: 1475–8.

631 https://doi.org/10.1098/rspb.2002.2040

632 Armanet N., L. Tosca, S. Brisset, T. Liehr, and G. Tachdjian, 2015 Small supernumerary marker

633 chromosomes in human infertility. Cytogen Genome Res 146: 100–108.

634 Bakkali M., and J. P. M. Camacho, 2004 The B chromosome polymorphism of the grasshopper

635 Eyprepocnemis plorans in North Africa. IV. Transmission of rare B chromosome variants. Cytogen

636 Genome Res 106: 332–7. https://doi.org/10.1159/000079308

637 Banaei-Moghaddam A. M., V. Schubert, K. Kumke, O. Weiβ, S. Klemme, et al., 2012 Nondisjunction

638 in favor of a chromosome: the mechanism of rye B chromosome drive during pollen mitosis. Plant

639 Cell 24: 4124–34. https://doi.org/10.1105/tpc.112.105270

640 Banaei-Moghaddam A. M., M. M. Martis, J. Macas, H. Gundlach, A. Himmelbach, et al., 2015 Genes

641 on B chromosomes: old questions revisited with new tools. Biochim Biophys Acta 1849: 64–70.

642 https://doi.org/10.1016/j.bbagrm.2014.11.007

643 Battaglia E., 1964 Cytogenetics of B-Chromosomes. Caryologia 17: 245–299.

644 https://doi.org/10.1080/00087114.1964.10796127 Page 33

645 Bauerly E., S. E. Hughes, D. R. Vietti, D. E. Miller, W. McDowell, et al., 2014 Discovery of

646 supernumerary B chromosomes in Drosophila melanogaster. Genetics 196: 1007–1016.

647 https://doi.org/10.1534/genetics.113.160556

648 Beckett J. B., 1991 Cytogenetic, genetic and plant breeding applications of B–A translocations in maize,

649 pp. 493–529 in Chromosome Engineering in Plants: Genetics, Breeding, Evolution (Part A), edited

650 by P. K. Gupta, T. Tsuchiya. Elsevier, Amsterdam.

651 Birchler J. A., 1991 Chromosome manipulations in maize, pp. 531–559 in Chromosome Engineering in

652 Plants: Genetics, Breeding, Evolution (Part A), edited by P. K. Gupta, T. Tsuchiya. Elsevier,

653 Amsterdam.

654 Bonner A. M., S. E. Hughes, J. A. Chisholm, S. K. Smith, B. D. Slaughter, et al., 2013 Binding of

655 Drosophila Polo kinase to its regulator Matrimony is noncanonical and involves two separate

656 functional domains. PNAS 110: E1222-31. https://doi.org/10.1073/pnas.1301690110

657 Breman A. M., F. J. Probst, M. A. Blazo, C. P. Schaaf, E. K. Roney, et al., 2011 Identification of

658 complex chromosome 18 rearrangements by FISH and array CGH in two patients with apparent

659 isochromosome 18q. Am J Med Genet A 155A: 1465–8. https://doi.org/10.1002/ajmg.a.33935

660 Bugge M., C. a Brandt, and M. B. Petersen, 2004 DNA studies of mono- and pseudodicentric

661 isochromosomes 18q. Am J Med Genet A 127A: 230–3. https://doi.org/10.1002/ajmg.a.30026

662 Camacho J., 2005 B chromosomes, pp. 223–286 in The Evolution of the Genome, edited by Gregory T.

663 R. Elsevier, San Diego.

664 Carlson W. R., 1970 Nondisjunction and isochromosome formation in the B chromosome of maize.

665 Chromosoma 30: 356–365. https://doi.org/10.1007/BF00321067

666 Cenci G., G. Siriaco, G. D. Raffa, R. Kellum, and M. Gatti, 2003 The Drosophila HOAP protein is Page 34

667 required for telomere capping. Nat Cell Biol 5: 82–4. https://doi.org/10.1038/ncb902

668 Clyde M., 1980 Chromosome IV variation in D. albomican Duda. Drosoph Inf Serv 55: 25–26.

669 D’Ambrosio U., M. P. Alonso-Lifante, K. Barros, A. Kovařík, G. Mas de Xaxars, et al., 2017 B-chrom:

670 a database on B-chromosomes of plants, animals and fungi. New Phytol 216: 635–642.

671 https://doi.org/10.1111/nph.14723

672 Darlington C. D., 1939 Misdivision and the genetics of the centromere. J Genet 37: 341–364.

673 https://doi.org/10.1007/BF02982733

674 de la Chapelle A., 1982 How do human isochromosomes arise? Cancer Genet Cytogenet 5: 173–9.

675 Deng Q., Q. Zeng, Y. Qian, C. Li, and Y. Yang, 2007 Research on the karyotype and evolution of

676 Drosophila melanogaster species group. J Genet Genomics 34: 196–213.

677 https://doi.org/10.1016/S1673-8527(07)60021-6

678 Dhar M. K., B. Friebe, A. K. Koul, and B. S. Gill, 2002 Origin of an apparent B chromosome by

679 mutation, chromosome fragmentation and specific DNA sequence amplification. Chromosoma 111:

680 332–40. https://doi.org/10.1007/s00412-002-0214-4

681 Dick C., and E. W. Johns, 1968 The effect of two acetic acid containing fixatives on the histone content

682 of calf thymus deoxyribonucleoprotein and calf thymus tissue. Exp Cell Res 51: 626–632.

683 https://doi.org/10.1016/0014-4827(68)90150-X

684 Ellis N. A., R. N. Douglas, C. E. Jackson, J. A. Birchler, and R. K. Dawe, 2015 Generation of a maize B

685 centromere minimal map containing the central core domain. G3 5: 2857–64.

686 https://doi.org/10.1534/g3.115.022889

687 Friebe B., P. Zhang, G. Linc, and B. S. Gill, 2005 Robertsonian translocations in wheat arise by centric

688 misdivision of univalents at anaphase I and rejoining of broken centromeres during interkinesis of Page 35

689 meiosis II. Cytogen Genome Res 109: 293–297. https://doi.org/10.1159/000082412

690 Fuster C., M. A. Rigola, and J. Egozcue, 2004 Human supernumeraries: are they B chromosomes?

691 Cytogen Genome Res 106: 165–172. https://doi.org/10.1159/000079283

692 Gao G., J.-C. Walser, M. L. Beaucher, P. Morciano, N. Wesolowska, et al., 2010 HipHop interacts with

693 HOAP and HP1 to protect Drosophila telomeres in a sequence-independent manner. EMBO J 29:

694 819–29. https://doi.org/10.1038/emboj.2009.394

695 Gutknecht J., D. Sperlich, and L. Bachmann, 1995 A species specific satellite DNA family of

696 Drosophila subsilvestris appearing predominantly in B chromosomes. Chromosoma 103: 539–44.

697 https://doi.org/10.1007/BF00355318

698 Han F., J. C. Lamb, W. Yu, Z. Gao, and J. A. Birchler, 2007 Centromere function and nondisjunction

699 are independent components of the maize B chromosome accumulation mechanism. Plant Cell 19:

700 524–33. https://doi.org/10.1105/tpc.106.049577

701 Han F., J. C. Lamb, M. E. McCaw, Z. Gao, B. Zhang, et al., 2018 Meiotic studies on combinations of

702 chromosomes with different sized centromeres in maize. Front Plant Sci 9: 785.

703 https://doi.org/10.3389/fpls.2018.00785

704 Harris D., C. Orme, J. Kramer, L. Namba, M. Champion, et al., 2003 A deficiency screen of the major

705 autosomes identifies a gene (matrimony) that is haplo-insufficient for achiasmate segregation in

706 Drosophila oocytes. Genetics 165: 637–52.

707 Huang W., Y. Du, X. Zhao, and W. Jin, 2016 B chromosome contains active genes and impacts the

708 transcription of A chromosomes in maize (Zea mays L.). BMC Plant Biol 16: 88.

709 https://doi.org/10.1186/s12870-016-0775-7

710 Jafari-Ghahfarokhi H., M. Moradi-Chaleshtori, T. Liehr, M. Hashemzadeh-Chaleshtori, H. Teimori, et Page 36

711 al., 2015 Small supernumerary marker chromosomes and their correlation with specific syndromes.

712 Adv Biomed Res 4: 140. https://doi.org/10.4103/2277-9175.161542

713 Johansen K. M., W. Cai, H. Deng, X. Bao, W. Zhang, et al., 2009 Polytene chromosome squash

714 methods for studying transcription and epigenetic chromatin modification in Drosophila using

715 antibodies. Methods 48: 387–397. https://doi.org/10.1016/j.ymeth.2009.02.019

716 John B., and G. M. Hewitt, 1965 The B-chromosome system of Myrmeleotettix maculatus (Thunb.).

717 Chromosoma 16: 548–578. https://doi.org/10.1007/BF00326973

718 Jones R. N., and H. Rees, 1982 B Chromosomes. Academic Press, London.

719 Jones R. N., 1995 B chromosomes in plants. New Phytol 131: 411–434. https://doi.org/10.2307/2558912

720 Kaszás E., and J. A. Birchler, 1996 Misdivision analysis of centromere structure in maize. EMBO J 15:

721 5246–55.

722 Kour G., S. Kaul, and M. K. Dhar, 2014 Molecular characterization of repetitive DNA sequences from

723 B chromosome in Plantago lagopus L. Cytogen Genome Res 142: 121–8.

724 https://doi.org/10.1159/000356472

725 Kurzhals R. L., L. Fanti, A. C. G. Ebsen, Y. S. Rong, S. Pimpinelli, et al., 2017 Chromosome healing is

726 promoted by the telomere cap component Hiphop in Drosophila. Genetics 207: 949–959.

727 https://doi.org/10.1534/genetics.117.300317

728 Lanzas P., F. Perfectti, M. A. Garrido-Ramos, C. Ruíz-Rejón, M. González-Sánchez, et al., 2018 Long-

729 term monitoring of B-chromosome invasion and neutralization in a population of Prospero

730 autumnale (Asparagaceae). Evolution 1–27. https://doi.org/10.1111/evo.13501

731 Larracuente A. M., and P. M. Ferree, 2015 Simple method for fluorescence DNA in situ hybridization to

732 squashed chromosomes. J Vis Exp 52288. https://doi.org/10.3791/52288 Page 37

733 Li H., and R. Durbin, 2009 Fast and accurate short read alignment with Burrows-Wheeler transform.

734 Bioinformatics 25: 1754–60. https://doi.org/10.1093/bioinformatics/btp324

735 Liehr T., and A. Weise, 2007 Frequency of small supernumerary marker chromosomes in prenatal,

736 newborn, developmentally retarded and infertility diagnostics. Int J Mol Med 19: 719–731.

737 Liehr T., K. Mrasek, N. Kosyakova, C. M. Ogilvie, J. Vermeesch, et al., 2008 Small supernumerary

738 marker chromosomes (sSMC) in humans; are there B chromosomes hidden among them. Mol

739 Cytogenet 1: 12. https://doi.org/10.1186/1755-8166-1-12

740 Liehr T., 2012 Small Supernumerary Marker Chromosomes (SMCs), A Guide for Human Geneticists

741 and Clinicians. Springer Berlin Heidelberg, Berlin, Heidelberg.

742 Litchfield K., M. Levy, R. A. Huddart, J. Shipley, and C. Turnbull, 2016 The genomic landscape of

743 testicular germ cell tumours: from susceptibility to treatment. Nat Rev Urol 13: 409–19.

744 https://doi.org/10.1038/nrurol.2016.107

745 Lohe A. R., A. J. Hilliker, and P. A. Roberts, 1993 Mapping simple repeated DNA sequences in

746 heterochromatin of Drosophila melanogaster. Genetics 134: 1149–74.

747 https://doi.org/10.1016/0168-9525(93)90135-5

748 López-León M. D., N. Neves, T. Schwarzacher, J. S. Heslop-Harrison, G. M. Hewitt, et al., 1994

749 Possible origin of a B chromosome deduced from its DNA composition using double FISH

750 technique. Chrom Res 2: 87–92. https://doi.org/10.1007/BF01553487

751 Lukaszewski A. J., 2010 Behavior of centromeres in univalents and centric misdivision in wheat.

752 Cytogen Genome Res 129: 97–109. https://doi.org/10.1159/000314108

753 Luo R., B. Liu, Y. Xie, Z. Li, W. Huang, et al., 2012 SOAPdenovo2: an empirically improved memory-

754 efficient short-read de novo assembler. Gigascience 1: 18. https://doi.org/10.1186/2047-217X-1-18 Page 38

755 Manrique-Poyato M. I., M. D. López-León, J. Cabrero, F. Perfectti, and J. P. M. Camacho, 2013 Spread

756 of a new parasitic B chromosome variant is facilitated by high gene flow. PLoS One 8: e83712.

757 https://doi.org/10.1371/journal.pone.0083712

758 Martis M. M., S. Klemme, A. M. Banaei-Moghaddam, F. R. Blattner, J. Macas, et al., 2012 Selfish

759 supernumerary chromosome reveals its origin as a mosaic of host genome and organellar

760 sequences. PNAS 109: 13343–6. https://doi.org/10.1073/pnas.1204237109

761 Masonbrink R. E., and J. A. Birchler, 2012 Accumulation of multiple copies of maize

762 minichromosomes. Cytogen Genome Res 137: 50–9. https://doi.org/10.1159/000339615

763 Mestriner C. A., P. M. Galetti, S. R. Valentini, I. R. Ruiz, L. D. Abel, et al., 2000 Structural and

764 functional evidence that a B chromosome in the characid fish Astyanax scabripinnis is an

765 isochromosome. Heredity 85 ( Pt 1): 1–9. https://doi.org/10.1046/j.1365-2540.2000.00702.x

766 Mitelman F., B. Johansson, and F. Mertens, 2018 Mitelman database of chromosome aberrations and

767 gene fusions in cancer. http://cgap.nci.nih.gov/Chromosomes/Mitelman.

768 Navarro-Domínguez B., F. J. Ruiz-Ruano, J. Cabrero, J. M. Corral, M. D. López-León, et al., 2017

769 Protein-coding genes in B chromosomes of the grasshopper Eyprepocnemis plorans. Sci Rep 7:

770 45200. https://doi.org/10.1038/srep45200

771 O’Leary N. A., M. W. Wright, J. R. Brister, S. Ciufo, D. Haddad, et al., 2016 Reference sequence

772 (RefSeq) database at NCBI: current status, taxonomic expansion, and functional annotation.

773 Nucleic Acids Res 44: D733–D745. https://doi.org/10.1093/nar/gkv1189

774 Pimpinelli S., S. Bonaccorsi, L. Fanti, and M. Gatti, 2011 Immunostaining of mitotic chromosomes from

775 Drosophila larval brain. CSH Prot 2011: pdb.prot065524-pdb.prot065524.

776 https://doi.org/10.1101/pdb.prot065524 Page 39

777 Poletto A. B., I. A. Ferreira, and C. Martins, 2010 The B chromosomes of the African cichlid fish

778 Haplochromis obliquidens harbour 18S rRNA gene copies. BMC Genet 11: 1.

779 https://doi.org/10.1186/1471-2156-11-1

780 Randolph L. F., 1928 Types of supernumerary chromosomes in maize. Anat Rec 41: 102.

781 Rhoades M. M., 1933 A secondary trisome in maize. PNAS 19: 1031–8.

782 https://doi.org/10.1073/pnas.19.12.1031

783 Ruban A., T. Schmutzer, U. Scholz, and A. Houben, 2017 How next-generation sequencing has aided

784 our understanding of the sequence composition and origin of B chromosomes. Genes (Basel) 8:

785 294. https://doi.org/10.3390/genes8110294

786 Ruiz-Ruano F. J., J. Cabrero, M. D. López-León, and J. P. M. Camacho, 2017 Satellite DNA content

787 illuminates the ancestry of a supernumerary (B) chromosome. Chromosoma 126: 487–500.

788 https://doi.org/10.1007/s00412-016-0611-8

789 Ruiz-Ruano F. J., J. Cabrero, M. D. López-León, A. Sánchez, and J. P. M. Camacho, 2018 Quantitative

790 sequence characterization for repetitive DNA content in the supernumerary chromosome of the

791 migratory locust. Chromosoma 127: 45–57. https://doi.org/10.1007/s00412-017-0644-7

792 Sears E. R., 1952 Misdivision of univalents in common wheat. Chromosoma 4: 535–50.

793 https://doi.org/10.1007/BF00325789

794 Smit A., R. Hubley, and P. Green, 2013-2015 RepeatMasker Open-4.0 .

795 Steinitz-Sears L. M., 1966 Somatic instability of telocentric chromosomes in wheat and the nature of the

796 centromere. Genetics 54: 241–8.

797 Sun M., H. Zhang, G. Li, C. J. Guy, X. Wang, et al., 2017 Molecular characterization of 20 small

798 supernumerary marker chromosome cases using array comparative genomic hybridization and Page 40

799 fluorescence in situ hybridization. Sci Rep 7: 10395. https://doi.org/10.1038/s41598-017-10466-z

800 Sundaran A. K., and J. P. Gupta, 1994 A new finding of B-chromosomes in Drosophila kikkawai Burla.

801 Rev Bras Genética 17: 223–224.

802 Talbert P., S. Kasinathan, and S. Henikoff, 2018 Simple and complex centromeric satellites in

803 Drosophila sibling species. Genetics genetics.300620.2017.

804 https://doi.org/10.1534/genetics.117.300620

805 Titen S. W. A., and K. G. Golic, 2008 Telomere loss provokes multiple pathways to apoptosis and

806 produces genomic instability in Drosophila melanogaster. Genetics 180: 1821–32.

807 https://doi.org/10.1534/genetics.108.093625

808 Tonomura Y., and Y. N. Tobari, 1983 Y-chromosome variation and B-chromosomes in Drosophila

809 malerkotliana. Sci Reports from Tokyo Women’s Christ Univ 32: 705–711.

810 Upcott M., 1937 The external mechanics of the chromosomes. VI. The behaviour of the centromere at

811 meiosis. Proc R Soc B Biol Sci 124: 336–361. https://doi.org/10.1098/rspb.1937.0089

812 Valente G. T., M. A. Conte, B. E. A. Fantinatti, D. C. Cabral-de-Mello, R. F. Carvalho, et al., 2014

813 Origin and evolution of B chromosomes in the cichlid fish Astatotilapia latifasciata based on

814 integrated genomic analyses. Mol Biol Evol 31: 2061–72. https://doi.org/10.1093/molbev/msu148

815 Valente G. T., R. T. Nakajima, B. E. A. Fantinatti, D. F. Marques, R. O. Almeida, et al., 2017 B

816 chromosomes: from cytogenetics to systems biology. Chromosoma 126: 73–81.

817 https://doi.org/10.1007/s00412-016-0613-6

818 Vicente V. E., O. Moreira-Filho, and J. P. Camacho, 1996 Sex-ratio distortion associated with the

819 presence of a B chromosome in Astyanax scabripinnis (Teleostei, Characidae). Cytogenet Cell

820 Genet 74: 70–5. https://doi.org/10.1159/000134385 Page 41

821 Wang L., Z. Xu, F. Iqbal, L. Zhong, Y. Zhang, et al., 2015 Decreased XY recombination and disturbed

822 meiotic prophase I progression in an infertile 48, XYY, +sSMC man. Chrom Res 23: 267–276.

823 https://doi.org/10.1007/s10577-015-9465-9

824 Wei K. H.-C., J. K. Grenier, D. A. Barbash, and A. G. Clark, 2014 Correlated variation and population

825 differentiation in satellite DNA abundance among lines of Drosophila melanogaster. PNAS 111:

826 201421951. https://doi.org/10.1073/pnas.1421951112

827 Wilson E. B., 1907 The supernumerary chromosomes of Hemiptera. Science 26: 870–871.

828 https://doi.org/10.1126/science.26.660.258

829 Xiang Y., S. Takeo, L. Florens, S. E. Hughes, L.-J. Huo, et al., 2007 The inhibition of polo kinase by

830 matrimony maintains G2 arrest in the meiotic cell cycle., (T. Orr-Weaver, Ed.). PLoS Biol 5: e323.

831 https://doi.org/10.1371/journal.pbio.0050323

832 Yu W., F. Han, Z. Gao, J. M. Vega, and J. A. Birchler, 2007 Construction and behavior of engineered

833 minichromosomes in maize. PNAS 104: 8924–9. https://doi.org/10.1073/pnas.0700932104

834 Zurita S., J. Cabrero, M. D. Lopez-Leon, and J. P. M. Camacho, 1998 Polymorphism regeneration for a

835 neutralized selfish B chromosome. Evolution 52: 274. https://doi.org/10.2307/2410945

836 Page 42

837 FIGURE LEGENDS

838 Figure 1 Cytological and molecular evaluation of the D. melanogaster B chromosomes. (A) Metaphase

126 839 chromosome spread of a male with 10 B chromosomes, collected from the mtrm stock that carries an

840 average of 10–12 B chromosomes. Scale bar = 5 μm. (B) Illustrated representation of the karyotype in

841 (A). (C) Results of PFGE after gel was stained with GelRed (image is inverted). The B chromosomes

842 form a single, distinct band around 1.81 Mbp that is absent in the wild-type (WT) stock. Size standard:

843 H. wingeii chromosomal DNA (Bio-Rad).

844

845 Figure 2 FISH on metaphase chromosome spreads using probes that recognize simple satellite repeat

846 sequences. (A) A probe recognizing the AATAT repeat hybridizes to the B chromosomes, Chromosome

847 4, the tip of the X chromosome, and various locations on the Y chromosome. Scale bar = 5 μm. (B) A

848 probe recognizing the AAGAT repeat hybridizes only to the B chromosomes and Chromosome 4,

849 indicating the B chromosomes arose from Chromosome 4. Scale bar = 5 μm. (C) Magnification of

850 Chromosome 4 (arrowhead) and the B chromosomes from the box in (B). Scale bar = 1 μm.

851

852 Figure 3 The B chromosomes have functional telomeres. Immunofluorescence with antibodies to CID

853 (present at centromeres) and HOAP (present at telomeres) on metaphase chromosome spreads with B

854 chromosomes. Magnified images are of Chromosome 4 (solid-line box in whole spread) and a B

855 chromosome (dashed-line box in whole spread). Scale bar in whole spreads = 5 μm; scale bar in

856 magnifications = 1 μm.

857

858 Figure 4 The B chromosomes have twice as much AAGAT sequence as Chromosome 4 as assayed by

859 FISH. Chromosome spreads showing under-condensed (A), moderately condensed (B), and well-

860 condensed (C) chromosomes. Magnified images of the AAGAT FISH signal on Chromosome 4 (in solid- Page 43

861 line boxes) and the signal on the B chromosomes (in dashed-line boxes) illustrate how B chromosomes

862 can have up to four AAGAT foci whereas Chromosome 4 has up to two. Scale bar in whole spreads = 5

863 μm; scale bar in magnifications = 1 μm. (D) Quantitative comparison of AAGAT FISH probe

864 fluorescence signal indicates a twofold intensity difference of the B chromosome when compared to

865 Chromosome 4.

866

867 Figure 5 The B2 chromosome carries the AAGAT repeat and has centromeres and telomeres. (A) A

868 FISH probe recognizing the AAGAT satellite repeat hybridizes to the B2 chromosome in metaphase

869 chromosome spreads. Inset shows the magnification of two B2 chromosomes contained within the

870 dashed-line box. Note: this larva carried three copies of Chromosome 4, which is not uncommon in a

871 mtrm null heterozygote (see text). (B) Immunofluorescence with antibodies to CID (present at

KG08051 872 centromeres) and HOAP (present at telomeres) on chromosome spreads in the mtrm stock that

873 carries the B2 chromosome. Inset shows the magnification of a B2 chromosome contained within the

874 dashed-line box. Scale bar in whole spreads = 5 μm; scale bar in magnifications (insets) = 0.5 μm.

875

876 Figure 6 The B1 and B2 chromosome variants are distinct. (A) Metaphase chromosome spread with

877 three copies of the B1 chromosome, two copies of the B2 chromosome, and a single copy of

878 Chromosome 4 (denoted by an arrowhead) Note: this larva carried one copy of Chromosome 4, which is

879 not uncommon in a mtrm null heterozygote (see text). Scale bar = 5 μm (B) Quantitative comparison of

880 AAGAT FISH probe fluorescence signal indicates the B2 chromosome has the same amount of signal as

881 Chromosome 4. (C) A cell in early anaphase was observed in the same tissue that produced (A). Here, as

882 in metaphase chromosome spreads, the B1 chromosome is observed to have twice the AAGAT signal as

883 the B2 chromosome. Scale bar = 5 μm (D) Quantitative comparison of DAPI staining intensity between

884 the B1 and B2 chromosome. indicating the B2 chromosome has less DNA content than the B1

885 chromosome. Page 44 KG08051 886 Figure 7 Other mtrm stocks do not carry B chromosomes and are polymorphic for the AAGAT

KG08051 887 satellite repeat. (A) The laboratory version of the mtrm stock carries the B2 chromosome and has

888 the AAGAT repeat on Chromosome 4 (denoted by arrowheads) as assessed by FISH. Note: this larva

889 carried three copies of Chromosome 4, which is not uncommon in a mtrm null heterozygote (see text).

KG08051 SIMR1 SIMR2 BDSC 890 The other mtrm lines mtrm (B), mtrm (C), and mtrm (D) do not possess B

891 chromosomes, nor do they carry the AAGAT satellite repeat (Chromosome 4 denoted by arrowheads).

892 Brightness and contrast were adjusted for the DNA (DAPI) channel only (see Materials and Methods).

893 Scale bar = 5 μm.

894

895 Figure 8 Model for B1 chromosome formation in D. melanogaster. FISH analysis of the AAGAT

896 repetitive sequence (depicted in yellow) is consistent with the structure of the B1 chromosome being an

897 isochromosome (see Figure 4 and text). Such a chromosome may form after the misdivision of the

898 centromere on Chromosome 4, leaving the fragments from the short, left arm to fuse and become the B1

899 chromosome. Table 1 List of the most abundant kmers from our sequence analysis of the B chromosome.

Kmer Count % of maximuma kmer category TAAGA 40,607,724 100% GATAA 31,022,996 76% ATAAG 27,013,024 67% AAGAT AAGAT 16,852,053 41% AGATA 16,545,657 41% TATAA 10,971,510 27% TAATA 10,414,421 26% AATAT 9,354,811 23% AATAT ATATA 7,010,495 17% ATAAT 4,812,487 12% A 4,235,133 10% A AGAAG 3,402,400 8% AAGAG 2,840,917 7% GAAGA 2,743,196 7% AAGAG AGAGA 2,652,461 7% GAGAA 2,172,940 5% a Refers to kmer abundance as compared to TAAGA, the highest frequency kmer observed.

A B C 126 2 WT mtrm 2 3.13 Mbp 4 4

Y Bs

2.7 Mbp 2.35 Mbp

1.81 Mbp B chr 3 X 1.66 Mbp 3 1.37 Mbp 1.05 Mbp A DNA AATAT DNA AATAT

B DNA AAGAT DNA AAGAT

C DNA AAGAT DNA AAGAT DNA CID HOAP DNA CID HOAP

Chr 4 B chr Chr 4 B chr Chr 4 B chr Chr 4 B chr C B A AAGAT AAGAT AAGAT DNA DNA DNA B chr B chr Chr 4 B chr Chr 4 Chr 4 D Fold difference of AAGAT intensity

0.0 (B0.5 chromosome/Chromosome1.0 1.5 2.0 2.5 4) 3.0 A DNA AAGAT DNA AAGAT

B DNA CID HOAP DNA CID HOAP A DNA AAGAT DNA AAGAT B2

B1 E DNA B 2.5 AC

2.0

1.5

1.0

0.5 AAGAT (B2 chromosome/Chromosome 4) Fold difference of AAGAT intensity 0.0

D 100%

80%

DNA 60% B1 B2 AAGAT

B1 B2 40% B1 B1 B2 B1 Percent DAPI intensity 20% B1 B2 (B2 chromosome/B1 Chromosome)

0% A DNA AAGAT AATAT DNA AAGAT AATAT

KG08051 B2 B2 B2 B2 mtrm

B SIMR1 mtrm

C SIMR2 mtrm

D BDSC mtrm AAGAT satellite repeat B1 in metaphase

Chromosome 4 Breakage in Centromere Fusion of broken short arms to attached to the centromeric region misdivision create an isochromosome spindle (proposed structure of B1 chromosome)